I'm building something to automate this. Copy paste recipe text/url -> get flowgraph.

Edit: thought the link was about this: https://fluidvisual.com/flowchartrecipes/

Edit 2: "Chapter 2 Resource theories: Monoidal preorders and enrichment", in particular page 2

https://ocw.mit.edu/courses/18-s097-applied-category-theory-...

Edit 3: one problem the representation used on www.cookingforengineers.com has is that it can only represent recipes whose structure is that of a tree: every step can have multiple direct ancestors but leads to only one descendant. Not all steps of all recipes have this "convergence" property. See the lemon meringue pie example in the link above, where yolk and whites are processed separately (filling and meringnue respectively) before being merged in the last step (covering the pie with meringue)

I'd like to see what you make! I see the value in the flow chart, especially the pie you linked.

Regarding the CfE multiple descendants problem, it's not really a problem.

    +-----+----------+-------+---...
    + Egg | separate | White | etc..
    +     |          |-------+---...
    +     |          | Yolk  | etc..
    +-----+----------+-------+---...

or you'd just start with "1 egg yolk" and "1 egg white" for your recipe. Or you'd say "See meringue below" or... some other answer that you like.

Whatever, we're humans, we roll with it.

If the recipe somehow gets too complicated, you make two separate tables, like they did for the recipe above.

My flowgraphs are not worth showing yet. Laying out graphs properly without human intervention is ... very hard, and this is a fascinating field with a lot of ongoing research.

I can however point you to some designs I generated using dall-e (non-sensical and with a perplexing flow, but the illustration are a good start towards something gorgeous): https://imgur.com/a/1NN8DaU

There are also other ways to use LLMs to explore culinary arts, for instance these "equations" pertaining to caramel:

    1. Creaminess (creamy texture in caramel)
    Creaminess ↑ ⇔ Cream ↑ + Butter ↑ + Milk ↑
    Creaminess ↓ ⇔ Sugar ↑ (hard caramel) + Water ↑ (syrup texture)
    
    2. Hardness (firmness or brittleness of the caramel)
    Hardness ↑ ⇔ Cooking temperature ↑ + Sugar ↑
    Hardness ↓ ⇔ Butter ↑ + Cream ↑ + Shorter cooking time ↓
    
    3. Chewiness (soft, stretchy caramel)
    Chewiness ↑ ⇔ Butter ↑ + Cream ↑ + Glucose syrup ↑
    Chewiness ↓ ⇔ Cooking temperature ↑ + Sugar ↑
    
    4. Stickiness (caramel that adheres to teeth or surfaces)
    Stickiness ↑ ⇔ Sugar ↑ + Longer cooking time ↑
    Stickiness ↓ ⇔ Butter ↑ + Cream ↑ + Shorter cooking time ↓
    
    5. Color (darker caramel)
    Color ↑ (darker) ⇔ Cooking temperature ↑ + Longer cooking time ↑
    Color ↓ (lighter) ⇔ Lower temperature ↓ + Shorter cooking time ↓
    
    6. Sweetness (perceived sugar taste)
    Sweetness ↑ ⇔ Sugar ↑ + Cooking time (slightly shorter) ↓
    Sweetness ↓ ⇔ Butter ↑ + Salt ↑ + Longer cooking time ↑ (more bitterness)
    
    7. Bitterness (burnt or deeper flavor)
    Bitterness ↑ ⇔ Cooking temperature ↑ + Longer cooking time ↑
    Bitterness ↓ ⇔ Lower temperature ↓ + Shorter cooking time ↓

It's difficult to get this kind of info without having a lot of experience in a specific culinary space. I've never cooked caramel seriously so I can't tell if this is right (looks like though), but I have started to master ice cream production this summer, and the set of equations ChatGPT generated was on point, so I guess I can trust those too.

Oh my god I love the way they visualize the recipe steps. That's how I write them down (grouping ingredients and steps), but its a table merging rows instead!

That is an absolutely brilliant way to visualize recipes.

Wow! Fantastic!